Horace Heffner <[email protected]> wrote:
> Two thermocouples were used to measure the delta T. If one is off 1 degree > hot absolute and the other is off 1 degree cold absolute, then delta T is > off by 2°C, systematically. > Yes of course if they are both off by 1 degree that's 2. HOWEVER: 1. They were not off by that much. Lewan and others saw they were 0.5°C apart. Someone should have turned the OFFSET adjustment to fix that bias. You can see that is the bias in the first several readings up to around 13:40, when Tout is always 0.5 to 0.7°C below Tin. The extra 0.2°C is probably caused by momentary fluctuations; they should have pressed the HOLD button before reading it. If they had done that, the bias would have been 0.5°C every time. 2. These thermocouples are in the same meter, so the precision will be the same even if there is a bias. 3. I have never heard of a laboratory grade 250 euro themometer with a 2°C bias at these temperatures, unless someone accidentally turned the OFFSET screw. That is a huge bias. 4. The inlet water is from the municipal reservoir which is gigantic. The temperature hardly varies. You can see this in inlet temperature record. It varies from 24.7 to 25.2°C. So you can ignore that and use only the outlet thermocouple change against itself, rather than the difference between them. Tout starts at 23.7°C, and then at the peak at 16:50 it is 35.5°C. It has moved up 11.8°C. Tin is 24.7°C at that time, so the difference Tout - Tin = 10.8°C. That is caused by the 0.5 deg C bias, and again probably by the fact they did not press HOLD so the temperature fluctuated before they could read both values at the same moment. It is very likely the actual difference is 11.8°C. But who cares? 10.8°C is fine. If there was not anomalous heat, the reactor would be stone cold by that time, and Tout would be 23.7°C. Why quibble about it? This is irrefutable proof. It makes no difference whether it was 7757 W or 8474 W. It sure as hell was not stone cold tap water temperature. > It looked to me the actual error was on the order of 0.8°C, but who knows > without calibration. > The first two hours before the steam entered the heat exchanger can be used as a calibration. > > It is possible, even easy, to make a thermocouple (pair) specifically for > measuring delta T. > Yes. This meter, the Termometro TM-947 SD has that feature built in. You just set it for T1-T2 mode. So does my HH12B. Of course when you set it to T1-T2 mode, you have to zero out the bias if there is one. I have never seen a bias as large as 0.5°C but if I did, I would just erase it with the OFFSET. I would also calibrate with cold water (5°C) tap water, and boiling water beforehand to make sure the thing is working. However, these things are extremely reliable and that is really only a formality. The likelihood that this instrument is wrong by 10°C and there is actually no heat is roughly as high as the likelihood that a new Prius digital speedometer will be off by ~10 mph. I'll bet that's never happened. I'll bet it *can't* happen. The thing might show zero or random numbers, but not a bias as large as 10 mph. > One means to greatly simplify and improve the delta T measurement is to > use paired type T thermocouples at the Tin and Tout measuring points. > These are paired. They are in the same meter. They share the same control electronics. Probably timeshare it. It is extremely unlikely they have significantly different precision at these temperatures; i.e. that one goes up 10.1°C while the other goes up 10.2°C. You do not need to take my word for any of this. If you do not believe my description of how these things work, I strongly recommend you buy or borrow a few of these things plus some conventional red liquid thermometers. Spend a day testing them at different temperatures. Be sure to stir the water. Don't use water with ice cubes in it. You will see how accurate and precise they are. You will see that there it not the slightest chance this one registered a 10°C difference when there was actually no difference. You can see from the final decay curve how quickly the thing cools. After 3 hours 1.8 tons of cooling water went through the system. Even if it had been incandescent before the power went off, it would have been cool by that time, and it was definitely not incandescent. The only possible explanation is that additional heat was generated after the power was turned off. Lots of power! More than 8 kW at times. - Jed
